Use the GDC 2021 Session Viewer to browse and sort the ever-growing list of sessions by pass type, topic, and format. You will be able to build your personal schedule directly within the event platform, once live early July 2021.
Shuo Xu (Lead AI Programmer, miHoYo Co., Ltd.)
Pass Type: All Access Pass, Core Pass
Vault Recording: TBD
Audience Level: N/A: I would like to deliver a virtual presentation only
This session reviews some of AI programming challenges in developing open-world game 'Genshin Impact' and describes several handful methods to improve AI system scalability. It details an AI framework and production pipeline that allow designers to craft huge amount of AI archetypes quickly and flexibly. Next, Genshin's pathfinding service is introduced, an approach to reduce NavMesh disk and memory space overhead, caused by continuously-growing world map. Finally, this session discusses AI performance optimization and shows how AI system can be modularized and broke down into multi-threaded tasks on CPU.
From this presentation attendees will learn techniques to make AI system more scalable by setting up AI pipelines and remote pathfinding services. Also, attendees will learn how to apply multi-threading for improving general AI performance.
This session is intended for programmers and designers who are interested in building open-world AI system. The audience is expected to have some basic knowledge of game AI development such as BehaviorTree, NavMesh, state machine, etc.